Face Generation¶

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data¶

You'll be using two datasets in this project:

MNIST
CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)

Found mnist Data
Found celeba Data

Explore the Data¶

MNIST¶

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')

<matplotlib.image.AxesImage at 0x15c39cf3d30>

CelebA¶

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))

<matplotlib.image.AxesImage at 0x15c39df83c8>

Preprocess the Data¶

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network¶

You'll build the components necessary to build a GANs by implementing the following functions below:

model_inputs
discriminator
generator
model_loss
model_opt
train

Check the Version of TensorFlow and Access to GPU¶

This will check to make sure you have the correct version of TensorFlow and access to a GPU

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))

C:\Users\deepai\Anaconda3\envs\u2\lib\site-packages\h5py\__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters

TensorFlow Version: 1.8.0
Default GPU Device: /device:GPU:0

Input¶

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
Z input placeholder with rank 2 using z_dim.
Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    # TODO: Implement Function
    real_input_images = tf.placeholder(tf.float32, [None, image_width,image_height,image_channels])
    tensor_of_z_data = tf.placeholder(tf.float32,[None, z_dim])
    learning_rate = tf.placeholder(tf.float32)
    
    return real_input_images, tensor_of_z_data, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)

Tests Passed

Discriminator¶

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the discriminator, tensor logits of the discriminator).

def discriminator(images, reuse=False):
    """
    Create the discriminator network
    :param images: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    # TODO: Implement Function
    #print("rimages", images.shape, "rreuse", reuse)
    with tf.variable_scope("discriminator", reuse=reuse):
        #     rconv1 (?, 14, 14, 64)
        #     rconv2 (?, 7, 7, 128)
        #     rconv3 (?, 4, 4, 256)  
        #     rlogits (?, 1)
        alpha = 0.1
        rconv1 = tf.layers.conv2d(images,64,3,strides=2,padding='same')
        rconv1 = tf.nn.leaky_relu(rconv1)
        #print("rconv1", rconv1.shape)
        
        rconv2 = tf.layers.conv2d(rconv1,128,3,strides=2,padding='same')
        rconv2 = tf.layers.batch_normalization(rconv2,training=True)
        rconv2 = tf.nn.leaky_relu(rconv2)
        #print("rconv2", rconv2.shape)

        rconv3 = tf.layers.conv2d(rconv2,256,2,strides=2,padding='same')
        rconv3 = tf.layers.batch_normalization(rconv3,training=True)
        rconv3 = tf.nn.leaky_relu(rconv3)
        #print("rconv3", rconv3.shape)

        
        rflattened = tf.reshape(rconv3,(-1,4*4*256))
        rlogits = tf.layers.dense(rflattened,1)
        rlogits = tf.layers.dropout(rlogits, rate=0.3)
        #print("rlogits", rlogits.shape)

        routput = tf.nn.sigmoid(rlogits)
        #print("output ", output.shape, "logits ",logits.shape)
    return routput, rlogits


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)

Tests Passed

Generator¶

Implement generator to generate an image using z. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

def generator(z, out_channel_dim, is_train=True):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    # TODO: Implement Function 
    #print("z", z.shape,"zout_channel_dim", out_channel_dim, "zis_train", is_train)
    with tf.variable_scope("generator", reuse=not is_train):
#         z (?, 100) out_channel_dim 5 is_train True
#         dense1 (?, 7, 7, 256)
#         zconv1 (?, 7, 7, 256)
#         zconv2 (?, 14, 14, 128)
#         zconv3 (?, 28, 28, 64)
#         logits (?, 28, 28, 5)
        alpha = 0.1
 
        zdense1 = tf.layers.dense(z,7*7*256)
        zdense1 = tf.reshape(zdense1,(-1,7,7,256))
        #print("zdense1", zdense1.shape)

        
        zconv1 = zdense1  
        zconv1 = tf.layers.batch_normalization(zconv1,training = is_train)
        zconv1 = tf.nn.leaky_relu(zconv1)
        #print("zconv1", zconv1.shape)

        
        zconv2 = tf.layers.conv2d_transpose(zconv1,128,3,strides=2,padding='same')
        zconv2 = tf.layers.batch_normalization(zconv2,training=is_train)
        zconv2 = tf.nn.leaky_relu(zconv2)
        #print("zconv2", zconv2.shape)

        zconv3 = tf.layers.conv2d_transpose(zconv2,64,3,strides=2,padding='same')
        zconv3 = tf.layers.batch_normalization(zconv3,training = is_train)
        zconv3 = tf.nn.leaky_relu(zconv3)
        #print("zconv3", zconv3.shape)

        zlogits = tf.layers.conv2d_transpose(zconv3,out_channel_dim,3,strides=1,padding='same')
        #print("zlogits", zlogits.shape)

        zout = tf.tanh(zlogits)
    return zout


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)

Tests Passed

Loss¶

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

discriminator(images, reuse=False)
generator(z, out_channel_dim, is_train=True)

def model_loss(input_real, input_z, out_channel_dim):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    # TODO: Implement Function
    
    smooth = 0.1
    g_model = generator(input_z, out_channel_dim, is_train=True)
    d_model_real, d_logits_real = discriminator(input_real)
    d_model_fake, d_logits_fake = discriminator(g_model, reuse=True)

    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, labels=tf.ones_like(d_model_real) * (1 - smooth)))
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.zeros_like(d_model_fake)))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.ones_like(d_model_fake)))

    d_loss = d_loss_real + d_loss_fake

    return d_loss, g_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)

Tests Passed

Optimization¶

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    # TODO: Implement Function
    
    t_vars = tf.trainable_variables()
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    g_vars = [var for var in t_vars if var.name.startswith('generator')]
    
    update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
    
    d_updates = [opt for opt in update_ops if opt.name.startswith('discriminator')]
    g_updates = [opt for opt in update_ops if opt.name.startswith('generator')]

    with tf.control_dependencies(d_updates):
        d_opt = tf.train.AdamOptimizer(learning_rate=learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)

    with tf.control_dependencies(g_updates):
        g_opt = tf.train.AdamOptimizer(learning_rate=learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)
            
    return d_opt, g_opt


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)

Tests Passed

Neural Network Training¶

Show Output¶

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train¶

Implement train to build and train the GANs. Use the following functions you implemented:

model_inputs(image_width, image_height, image_channels, z_dim)
model_loss(input_real, input_z, out_channel_dim)
model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    # TODO: Build Model
    


    print_every=10
    show_every=200
    losses = []
    n_images_to_show=20
    
    real_inputs, fake_inputs, lr_tensor  = model_inputs(data_shape[1], data_shape[2], data_shape[3], z_dim)
    d_loss, g_loss = model_loss(real_inputs, fake_inputs, data_shape[3])
    d_opt, g_opt = model_opt(d_loss, g_loss, learning_rate, beta1)

    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        n_steps=sum(1 for x in get_batches(batch_size))
        for counter_epoch in range(epoch_count): 
            print("STARTING epoch:",counter_epoch,", steps:", n_steps, "........")
            for step, img_batch_real in enumerate(get_batches(batch_size)):
                img_batch_real = img_batch_real*2.0
                # G:random noise 
                img_batch_fake = np.random.uniform(-1, 1, size=(batch_size, z_dim))
                
                # Run optimizers
                sess.run(d_opt, feed_dict={real_inputs: img_batch_real, 
                                                 fake_inputs: img_batch_fake, 
                                                 lr_tensor: learning_rate})
                
                sess.run(g_opt, feed_dict={real_inputs: img_batch_real, 
                                                 fake_inputs: img_batch_fake, 
                                                 lr_tensor: learning_rate})
                
                if step % print_every == 0: 
                    percent_step = (step / n_steps)*100
                    # At the end of each epoch, get the losses and print them out
                    train_loss_d = d_loss.eval({real_inputs: img_batch_real, fake_inputs: img_batch_fake})
                    train_loss_g = g_loss.eval({fake_inputs: img_batch_fake})
                    #run it twice http://bamos.github.io/2016/08/09/deep-completion/
                    train_loss_g = g_loss.eval({fake_inputs: img_batch_fake})

                    print("Epoch {}/{}...progress:{:.2f}%....step:{}/{}...".format(counter_epoch+1, epochs, percent_step, step,n_steps),
                          "Discriminator Loss: {:.4f}...".format(train_loss_d),
                          "Generator Loss: {:.4f}".format(train_loss_g))
                    # Save losses to view after training
                    losses.append((train_loss_d, train_loss_g))
                
                if step % show_every == 0:
                    show_generator_output(sess, n_images_to_show, fake_inputs, data_shape[3], data_image_mode)
        show_generator_output(sess, n_images_to_show, fake_inputs, data_shape[3], data_image_mode)
        print("Last train discriminator Loss: {:.4f}\nLast train generator Loss: {:.4f}".format(losses[-1][0], losses[-1][1]))

MNIST¶

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

batch_size = 64
z_dim = 200
learning_rate = 0.0005
beta1 = 0.1
#epochs = 25



"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 25

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)

STARTING epoch: 0 , steps: 937 ........
Epoch 1/25...progress:0.00%....step:0/937... Discriminator Loss: 5.6959... Generator Loss: 0.0080

Epoch 1/25...progress:1.07%....step:10/937... Discriminator Loss: 3.3735... Generator Loss: 0.0676
Epoch 1/25...progress:2.13%....step:20/937... Discriminator Loss: 3.6612... Generator Loss: 0.0401
Epoch 1/25...progress:3.20%....step:30/937... Discriminator Loss: 2.5097... Generator Loss: 0.1376
Epoch 1/25...progress:4.27%....step:40/937... Discriminator Loss: 2.2156... Generator Loss: 0.1899
Epoch 1/25...progress:5.34%....step:50/937... Discriminator Loss: 2.0229... Generator Loss: 0.2477
Epoch 1/25...progress:6.40%....step:60/937... Discriminator Loss: 1.9186... Generator Loss: 0.2745
Epoch 1/25...progress:7.47%....step:70/937... Discriminator Loss: 2.0419... Generator Loss: 0.2259
Epoch 1/25...progress:8.54%....step:80/937... Discriminator Loss: 1.6845... Generator Loss: 0.3639
Epoch 1/25...progress:9.61%....step:90/937... Discriminator Loss: 1.9078... Generator Loss: 0.2769
Epoch 1/25...progress:10.67%....step:100/937... Discriminator Loss: 1.4096... Generator Loss: 0.6966
Epoch 1/25...progress:11.74%....step:110/937... Discriminator Loss: 1.3565... Generator Loss: 1.1402
Epoch 1/25...progress:12.81%....step:120/937... Discriminator Loss: 1.4128... Generator Loss: 0.6194
Epoch 1/25...progress:13.87%....step:130/937... Discriminator Loss: 1.4459... Generator Loss: 0.5736
Epoch 1/25...progress:14.94%....step:140/937... Discriminator Loss: 1.6847... Generator Loss: 0.3360
Epoch 1/25...progress:16.01%....step:150/937... Discriminator Loss: 1.5289... Generator Loss: 0.4594
Epoch 1/25...progress:17.08%....step:160/937... Discriminator Loss: 1.4220... Generator Loss: 0.5226
Epoch 1/25...progress:18.14%....step:170/937... Discriminator Loss: 1.7239... Generator Loss: 0.3301
Epoch 1/25...progress:19.21%....step:180/937... Discriminator Loss: 1.4400... Generator Loss: 0.4541
Epoch 1/25...progress:20.28%....step:190/937... Discriminator Loss: 1.2653... Generator Loss: 0.6670
Epoch 1/25...progress:21.34%....step:200/937... Discriminator Loss: 1.2651... Generator Loss: 1.4274

Epoch 1/25...progress:22.41%....step:210/937... Discriminator Loss: 1.1326... Generator Loss: 1.0597
Epoch 1/25...progress:23.48%....step:220/937... Discriminator Loss: 1.4462... Generator Loss: 1.2798
Epoch 1/25...progress:24.55%....step:230/937... Discriminator Loss: 1.2068... Generator Loss: 1.3516
Epoch 1/25...progress:25.61%....step:240/937... Discriminator Loss: 1.2249... Generator Loss: 1.2480
Epoch 1/25...progress:26.68%....step:250/937... Discriminator Loss: 1.2182... Generator Loss: 0.8471
Epoch 1/25...progress:27.75%....step:260/937... Discriminator Loss: 1.1555... Generator Loss: 1.3633
Epoch 1/25...progress:28.82%....step:270/937... Discriminator Loss: 1.2213... Generator Loss: 1.0607
Epoch 1/25...progress:29.88%....step:280/937... Discriminator Loss: 1.2657... Generator Loss: 1.5607
Epoch 1/25...progress:30.95%....step:290/937... Discriminator Loss: 1.1959... Generator Loss: 1.4690
Epoch 1/25...progress:32.02%....step:300/937... Discriminator Loss: 1.1427... Generator Loss: 0.9956
Epoch 1/25...progress:33.08%....step:310/937... Discriminator Loss: 1.2805... Generator Loss: 1.2389
Epoch 1/25...progress:34.15%....step:320/937... Discriminator Loss: 1.2677... Generator Loss: 1.1005
Epoch 1/25...progress:35.22%....step:330/937... Discriminator Loss: 1.3308... Generator Loss: 0.7009
Epoch 1/25...progress:36.29%....step:340/937... Discriminator Loss: 1.8115... Generator Loss: 0.2875
Epoch 1/25...progress:37.35%....step:350/937... Discriminator Loss: 1.2943... Generator Loss: 0.6457
Epoch 1/25...progress:38.42%....step:360/937... Discriminator Loss: 1.3196... Generator Loss: 1.0956
Epoch 1/25...progress:39.49%....step:370/937... Discriminator Loss: 1.2558... Generator Loss: 1.0755
Epoch 1/25...progress:40.55%....step:380/937... Discriminator Loss: 1.0546... Generator Loss: 1.2528
Epoch 1/25...progress:41.62%....step:390/937... Discriminator Loss: 1.3208... Generator Loss: 1.1179
Epoch 1/25...progress:42.69%....step:400/937... Discriminator Loss: 1.3622... Generator Loss: 0.9004

Epoch 1/25...progress:43.76%....step:410/937... Discriminator Loss: 1.1957... Generator Loss: 0.9969
Epoch 1/25...progress:44.82%....step:420/937... Discriminator Loss: 1.2030... Generator Loss: 0.9835
Epoch 1/25...progress:45.89%....step:430/937... Discriminator Loss: 1.3444... Generator Loss: 0.8093
Epoch 1/25...progress:46.96%....step:440/937... Discriminator Loss: 1.2898... Generator Loss: 1.4431
Epoch 1/25...progress:48.03%....step:450/937... Discriminator Loss: 1.5418... Generator Loss: 0.4208
Epoch 1/25...progress:49.09%....step:460/937... Discriminator Loss: 1.6447... Generator Loss: 0.3540
Epoch 1/25...progress:50.16%....step:470/937... Discriminator Loss: 1.8748... Generator Loss: 0.2920
Epoch 1/25...progress:51.23%....step:480/937... Discriminator Loss: 1.1960... Generator Loss: 0.8502
Epoch 1/25...progress:52.29%....step:490/937... Discriminator Loss: 1.2485... Generator Loss: 0.7397
Epoch 1/25...progress:53.36%....step:500/937... Discriminator Loss: 1.3224... Generator Loss: 1.2017
Epoch 1/25...progress:54.43%....step:510/937... Discriminator Loss: 1.3414... Generator Loss: 1.1961
Epoch 1/25...progress:55.50%....step:520/937... Discriminator Loss: 1.2135... Generator Loss: 0.7493
Epoch 1/25...progress:56.56%....step:530/937... Discriminator Loss: 1.2552... Generator Loss: 0.7056
Epoch 1/25...progress:57.63%....step:540/937... Discriminator Loss: 1.7683... Generator Loss: 0.3170
Epoch 1/25...progress:58.70%....step:550/937... Discriminator Loss: 1.4715... Generator Loss: 0.4612
Epoch 1/25...progress:59.77%....step:560/937... Discriminator Loss: 1.6511... Generator Loss: 0.3833
Epoch 1/25...progress:60.83%....step:570/937... Discriminator Loss: 1.1548... Generator Loss: 0.9937
Epoch 1/25...progress:61.90%....step:580/937... Discriminator Loss: 1.1159... Generator Loss: 0.9279
Epoch 1/25...progress:62.97%....step:590/937... Discriminator Loss: 1.1554... Generator Loss: 0.9561
Epoch 1/25...progress:64.03%....step:600/937... Discriminator Loss: 1.1628... Generator Loss: 1.4826

Epoch 1/25...progress:65.10%....step:610/937... Discriminator Loss: 1.1247... Generator Loss: 1.2818
Epoch 1/25...progress:66.17%....step:620/937... Discriminator Loss: 1.3354... Generator Loss: 0.5425
Epoch 1/25...progress:67.24%....step:630/937... Discriminator Loss: 1.1652... Generator Loss: 1.3732
Epoch 1/25...progress:68.30%....step:640/937... Discriminator Loss: 1.3576... Generator Loss: 1.4561
Epoch 1/25...progress:69.37%....step:650/937... Discriminator Loss: 1.1809... Generator Loss: 0.7860
Epoch 1/25...progress:70.44%....step:660/937... Discriminator Loss: 1.5806... Generator Loss: 0.3845
Epoch 1/25...progress:71.50%....step:670/937... Discriminator Loss: 1.4439... Generator Loss: 0.4601
Epoch 1/25...progress:72.57%....step:680/937... Discriminator Loss: 1.0374... Generator Loss: 1.0947
Epoch 1/25...progress:73.64%....step:690/937... Discriminator Loss: 0.9474... Generator Loss: 1.0454
Epoch 1/25...progress:74.71%....step:700/937... Discriminator Loss: 1.5626... Generator Loss: 1.9442
Epoch 1/25...progress:75.77%....step:710/937... Discriminator Loss: 1.1058... Generator Loss: 1.3086
Epoch 1/25...progress:76.84%....step:720/937... Discriminator Loss: 1.2706... Generator Loss: 0.5790
Epoch 1/25...progress:77.91%....step:730/937... Discriminator Loss: 1.2343... Generator Loss: 0.6178
Epoch 1/25...progress:78.98%....step:740/937... Discriminator Loss: 1.0599... Generator Loss: 1.6570
Epoch 1/25...progress:80.04%....step:750/937... Discriminator Loss: 1.1958... Generator Loss: 0.8905
Epoch 1/25...progress:81.11%....step:760/937... Discriminator Loss: 1.0207... Generator Loss: 0.8548
Epoch 1/25...progress:82.18%....step:770/937... Discriminator Loss: 1.0975... Generator Loss: 0.9480
Epoch 1/25...progress:83.24%....step:780/937... Discriminator Loss: 0.9300... Generator Loss: 0.9415
Epoch 1/25...progress:84.31%....step:790/937... Discriminator Loss: 1.4328... Generator Loss: 0.5564
Epoch 1/25...progress:85.38%....step:800/937... Discriminator Loss: 1.2954... Generator Loss: 0.5618

Epoch 1/25...progress:86.45%....step:810/937... Discriminator Loss: 1.2081... Generator Loss: 0.6357
Epoch 1/25...progress:87.51%....step:820/937... Discriminator Loss: 0.9908... Generator Loss: 1.0606
Epoch 1/25...progress:88.58%....step:830/937... Discriminator Loss: 1.0706... Generator Loss: 1.7000
Epoch 1/25...progress:89.65%....step:840/937... Discriminator Loss: 1.9505... Generator Loss: 1.8927
Epoch 1/25...progress:90.72%....step:850/937... Discriminator Loss: 0.9565... Generator Loss: 1.0164
Epoch 1/25...progress:91.78%....step:860/937... Discriminator Loss: 1.1485... Generator Loss: 1.9571
Epoch 1/25...progress:92.85%....step:870/937... Discriminator Loss: 1.3340... Generator Loss: 0.5233
Epoch 1/25...progress:93.92%....step:880/937... Discriminator Loss: 1.0215... Generator Loss: 0.8186
Epoch 1/25...progress:94.98%....step:890/937... Discriminator Loss: 0.8656... Generator Loss: 1.0711
Epoch 1/25...progress:96.05%....step:900/937... Discriminator Loss: 1.2737... Generator Loss: 0.6120
Epoch 1/25...progress:97.12%....step:910/937... Discriminator Loss: 0.8482... Generator Loss: 1.4715
Epoch 1/25...progress:98.19%....step:920/937... Discriminator Loss: 1.1320... Generator Loss: 1.3715
Epoch 1/25...progress:99.25%....step:930/937... Discriminator Loss: 0.9499... Generator Loss: 1.0532
STARTING epoch: 1 , steps: 937 ........
Epoch 2/25...progress:0.00%....step:0/937... Discriminator Loss: 1.5063... Generator Loss: 0.4368

Epoch 2/25...progress:1.07%....step:10/937... Discriminator Loss: 0.8142... Generator Loss: 1.1417
Epoch 2/25...progress:2.13%....step:20/937... Discriminator Loss: 0.8526... Generator Loss: 1.1372
Epoch 2/25...progress:3.20%....step:30/937... Discriminator Loss: 1.6947... Generator Loss: 0.3727
Epoch 2/25...progress:4.27%....step:40/937... Discriminator Loss: 1.7169... Generator Loss: 0.3478
Epoch 2/25...progress:5.34%....step:50/937... Discriminator Loss: 1.1435... Generator Loss: 0.7117
Epoch 2/25...progress:6.40%....step:60/937... Discriminator Loss: 1.0801... Generator Loss: 1.4584
Epoch 2/25...progress:7.47%....step:70/937... Discriminator Loss: 1.2051... Generator Loss: 0.6827
Epoch 2/25...progress:8.54%....step:80/937... Discriminator Loss: 0.9125... Generator Loss: 1.2656
Epoch 2/25...progress:9.61%....step:90/937... Discriminator Loss: 1.1060... Generator Loss: 0.7251
Epoch 2/25...progress:10.67%....step:100/937... Discriminator Loss: 1.2935... Generator Loss: 2.3245
Epoch 2/25...progress:11.74%....step:110/937... Discriminator Loss: 0.9599... Generator Loss: 1.4033
Epoch 2/25...progress:12.81%....step:120/937... Discriminator Loss: 0.9325... Generator Loss: 0.9538
Epoch 2/25...progress:13.87%....step:130/937... Discriminator Loss: 1.5049... Generator Loss: 0.4595
Epoch 2/25...progress:14.94%....step:140/937... Discriminator Loss: 0.9373... Generator Loss: 0.9545
Epoch 2/25...progress:16.01%....step:150/937... Discriminator Loss: 1.4540... Generator Loss: 1.6010
Epoch 2/25...progress:17.08%....step:160/937... Discriminator Loss: 1.0288... Generator Loss: 1.1168
Epoch 2/25...progress:18.14%....step:170/937... Discriminator Loss: 1.1122... Generator Loss: 0.7980
Epoch 2/25...progress:19.21%....step:180/937... Discriminator Loss: 0.9393... Generator Loss: 1.0022
Epoch 2/25...progress:20.28%....step:190/937... Discriminator Loss: 1.0572... Generator Loss: 0.8128
Epoch 2/25...progress:21.34%....step:200/937... Discriminator Loss: 0.8507... Generator Loss: 1.1894

Epoch 2/25...progress:22.41%....step:210/937... Discriminator Loss: 1.7807... Generator Loss: 0.4325
Epoch 2/25...progress:23.48%....step:220/937... Discriminator Loss: 1.8329... Generator Loss: 0.2985
Epoch 2/25...progress:24.55%....step:230/937... Discriminator Loss: 0.9650... Generator Loss: 1.0532
Epoch 2/25...progress:25.61%....step:240/937... Discriminator Loss: 0.8787... Generator Loss: 1.1139
Epoch 2/25...progress:26.68%....step:250/937... Discriminator Loss: 1.0581... Generator Loss: 0.9184
Epoch 2/25...progress:27.75%....step:260/937... Discriminator Loss: 1.6335... Generator Loss: 0.3888
Epoch 2/25...progress:28.82%....step:270/937... Discriminator Loss: 1.5138... Generator Loss: 0.4858
Epoch 2/25...progress:29.88%....step:280/937... Discriminator Loss: 0.7792... Generator Loss: 1.4907
Epoch 2/25...progress:30.95%....step:290/937... Discriminator Loss: 0.8625... Generator Loss: 1.1765
Epoch 2/25...progress:32.02%....step:300/937... Discriminator Loss: 0.9514... Generator Loss: 0.9410
Epoch 2/25...progress:33.08%....step:310/937... Discriminator Loss: 1.2996... Generator Loss: 2.9156
Epoch 2/25...progress:34.15%....step:320/937... Discriminator Loss: 0.8811... Generator Loss: 1.3916
Epoch 2/25...progress:35.22%....step:330/937... Discriminator Loss: 0.9687... Generator Loss: 0.9508
Epoch 2/25...progress:36.29%....step:340/937... Discriminator Loss: 0.9955... Generator Loss: 0.9292
Epoch 2/25...progress:37.35%....step:350/937... Discriminator Loss: 1.0894... Generator Loss: 0.7615
Epoch 2/25...progress:38.42%....step:360/937... Discriminator Loss: 1.7441... Generator Loss: 0.3395
Epoch 2/25...progress:39.49%....step:370/937... Discriminator Loss: 1.1510... Generator Loss: 0.7876
Epoch 2/25...progress:40.55%....step:380/937... Discriminator Loss: 1.2726... Generator Loss: 0.6053
Epoch 2/25...progress:41.62%....step:390/937... Discriminator Loss: 1.2653... Generator Loss: 0.5734
Epoch 2/25...progress:42.69%....step:400/937... Discriminator Loss: 0.7829... Generator Loss: 1.6051

Epoch 2/25...progress:43.76%....step:410/937... Discriminator Loss: 0.7814... Generator Loss: 1.1891
Epoch 2/25...progress:44.82%....step:420/937... Discriminator Loss: 0.8847... Generator Loss: 1.0404
Epoch 2/25...progress:45.89%....step:430/937... Discriminator Loss: 2.1620... Generator Loss: 0.2456
Epoch 2/25...progress:46.96%....step:440/937... Discriminator Loss: 1.4828... Generator Loss: 0.4399
Epoch 2/25...progress:48.03%....step:450/937... Discriminator Loss: 1.5082... Generator Loss: 0.4734
Epoch 2/25...progress:49.09%....step:460/937... Discriminator Loss: 1.0412... Generator Loss: 1.2437
Epoch 2/25...progress:50.16%....step:470/937... Discriminator Loss: 1.3228... Generator Loss: 0.8736
Epoch 2/25...progress:51.23%....step:480/937... Discriminator Loss: 1.0464... Generator Loss: 0.8714
Epoch 2/25...progress:52.29%....step:490/937... Discriminator Loss: 1.0945... Generator Loss: 0.7886
Epoch 2/25...progress:53.36%....step:500/937... Discriminator Loss: 1.6366... Generator Loss: 0.3862
Epoch 2/25...progress:54.43%....step:510/937... Discriminator Loss: 1.6855... Generator Loss: 0.3763
Epoch 2/25...progress:55.50%....step:520/937... Discriminator Loss: 1.3879... Generator Loss: 0.4909
Epoch 2/25...progress:56.56%....step:530/937... Discriminator Loss: 1.0345... Generator Loss: 0.8158
Epoch 2/25...progress:57.63%....step:540/937... Discriminator Loss: 1.2578... Generator Loss: 0.6044
Epoch 2/25...progress:58.70%....step:550/937... Discriminator Loss: 1.4787... Generator Loss: 0.4501
Epoch 2/25...progress:59.77%....step:560/937... Discriminator Loss: 1.1213... Generator Loss: 1.7758
Epoch 2/25...progress:60.83%....step:570/937... Discriminator Loss: 1.1744... Generator Loss: 1.6401
Epoch 2/25...progress:61.90%....step:580/937... Discriminator Loss: 0.9554... Generator Loss: 1.1442
Epoch 2/25...progress:62.97%....step:590/937... Discriminator Loss: 0.9890... Generator Loss: 0.9000
Epoch 2/25...progress:64.03%....step:600/937... Discriminator Loss: 0.7886... Generator Loss: 1.3134

Epoch 2/25...progress:65.10%....step:610/937... Discriminator Loss: 1.2327... Generator Loss: 2.8328
Epoch 2/25...progress:66.17%....step:620/937... Discriminator Loss: 0.9134... Generator Loss: 1.0148
Epoch 2/25...progress:67.24%....step:630/937... Discriminator Loss: 1.0210... Generator Loss: 1.5658
Epoch 2/25...progress:68.30%....step:640/937... Discriminator Loss: 1.2377... Generator Loss: 1.8374
Epoch 2/25...progress:69.37%....step:650/937... Discriminator Loss: 1.1628... Generator Loss: 0.7841
Epoch 2/25...progress:70.44%....step:660/937... Discriminator Loss: 1.1072... Generator Loss: 1.2381
Epoch 2/25...progress:71.50%....step:670/937... Discriminator Loss: 1.1794... Generator Loss: 1.8041
Epoch 2/25...progress:72.57%....step:680/937... Discriminator Loss: 1.0088... Generator Loss: 1.4264
Epoch 2/25...progress:73.64%....step:690/937... Discriminator Loss: 0.8217... Generator Loss: 1.2710
Epoch 2/25...progress:74.71%....step:700/937... Discriminator Loss: 1.1296... Generator Loss: 0.7516
Epoch 2/25...progress:75.77%....step:710/937... Discriminator Loss: 1.2249... Generator Loss: 0.7062
Epoch 2/25...progress:76.84%....step:720/937... Discriminator Loss: 1.0634... Generator Loss: 1.5033
Epoch 2/25...progress:77.91%....step:730/937... Discriminator Loss: 0.8853... Generator Loss: 1.5371
Epoch 2/25...progress:78.98%....step:740/937... Discriminator Loss: 0.8817... Generator Loss: 1.0291
Epoch 2/25...progress:80.04%....step:750/937... Discriminator Loss: 0.9861... Generator Loss: 0.8836
Epoch 2/25...progress:81.11%....step:760/937... Discriminator Loss: 0.6500... Generator Loss: 1.8775
Epoch 2/25...progress:82.18%....step:770/937... Discriminator Loss: 1.1927... Generator Loss: 0.6653
Epoch 2/25...progress:83.24%....step:780/937... Discriminator Loss: 1.4274... Generator Loss: 0.4767
Epoch 2/25...progress:84.31%....step:790/937... Discriminator Loss: 1.3179... Generator Loss: 0.5701
Epoch 2/25...progress:85.38%....step:800/937... Discriminator Loss: 1.7780... Generator Loss: 0.3093

CelebA¶

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

batch_size = 32 #64
z_dim = 200
learning_rate = 0.0005
beta1 = 0.1
#epochs = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 25

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)

STARTING epoch: 0 , steps: 3780 ........
Epoch 1/25...progress:0.00%....step:0/3780... Discriminator Loss: 8.7570... Generator Loss: 0.0005

Epoch 1/25...progress:0.26%....step:10/3780... Discriminator Loss: 5.0758... Generator Loss: 0.0127
Epoch 1/25...progress:0.53%....step:20/3780... Discriminator Loss: 3.6278... Generator Loss: 0.0493
Epoch 1/25...progress:0.79%....step:30/3780... Discriminator Loss: 4.9574... Generator Loss: 0.0131
Epoch 1/25...progress:1.06%....step:40/3780... Discriminator Loss: 3.1103... Generator Loss: 0.0888
Epoch 1/25...progress:1.32%....step:50/3780... Discriminator Loss: 3.1505... Generator Loss: 0.0748
Epoch 1/25...progress:1.59%....step:60/3780... Discriminator Loss: 2.1760... Generator Loss: 0.2407
Epoch 1/25...progress:1.85%....step:70/3780... Discriminator Loss: 0.9888... Generator Loss: 1.1586
Epoch 1/25...progress:2.12%....step:80/3780... Discriminator Loss: 1.6730... Generator Loss: 1.4916
Epoch 1/25...progress:2.38%....step:90/3780... Discriminator Loss: 2.0094... Generator Loss: 3.3070
Epoch 1/25...progress:2.65%....step:100/3780... Discriminator Loss: 1.5014... Generator Loss: 2.6742
Epoch 1/25...progress:2.91%....step:110/3780... Discriminator Loss: 1.5023... Generator Loss: 1.9554
Epoch 1/25...progress:3.17%....step:120/3780... Discriminator Loss: 0.7261... Generator Loss: 2.0971
Epoch 1/25...progress:3.44%....step:130/3780... Discriminator Loss: 1.5347... Generator Loss: 2.8001
Epoch 1/25...progress:3.70%....step:140/3780... Discriminator Loss: 1.0502... Generator Loss: 2.3297
Epoch 1/25...progress:3.97%....step:150/3780... Discriminator Loss: 1.4371... Generator Loss: 2.4099
Epoch 1/25...progress:4.23%....step:160/3780... Discriminator Loss: 1.1832... Generator Loss: 2.1456
Epoch 1/25...progress:4.50%....step:170/3780... Discriminator Loss: 1.1026... Generator Loss: 2.3107
Epoch 1/25...progress:4.76%....step:180/3780... Discriminator Loss: 1.3216... Generator Loss: 1.8509
Epoch 1/25...progress:5.03%....step:190/3780... Discriminator Loss: 2.2454... Generator Loss: 0.1726
Epoch 1/25...progress:5.29%....step:200/3780... Discriminator Loss: 0.7252... Generator Loss: 1.7639

Epoch 1/25...progress:5.56%....step:210/3780... Discriminator Loss: 1.7747... Generator Loss: 1.0451
Epoch 1/25...progress:5.82%....step:220/3780... Discriminator Loss: 0.9540... Generator Loss: 1.6863
Epoch 1/25...progress:6.08%....step:230/3780... Discriminator Loss: 0.9758... Generator Loss: 1.9432
Epoch 1/25...progress:6.35%....step:240/3780... Discriminator Loss: 1.0862... Generator Loss: 2.6369
Epoch 1/25...progress:6.61%....step:250/3780... Discriminator Loss: 0.6774... Generator Loss: 2.2326
Epoch 1/25...progress:6.88%....step:260/3780... Discriminator Loss: 1.6624... Generator Loss: 2.1989
Epoch 1/25...progress:7.14%....step:270/3780... Discriminator Loss: 1.2372... Generator Loss: 1.0685
Epoch 1/25...progress:7.41%....step:280/3780... Discriminator Loss: 1.7065... Generator Loss: 0.6369
Epoch 1/25...progress:7.67%....step:290/3780... Discriminator Loss: 0.8377... Generator Loss: 1.5820
Epoch 1/25...progress:7.94%....step:300/3780... Discriminator Loss: 0.8913... Generator Loss: 1.6624
Epoch 1/25...progress:8.20%....step:310/3780... Discriminator Loss: 1.4994... Generator Loss: 0.8040
Epoch 1/25...progress:8.47%....step:320/3780... Discriminator Loss: 1.5464... Generator Loss: 2.4708
Epoch 1/25...progress:8.73%....step:330/3780... Discriminator Loss: 1.0927... Generator Loss: 0.8450
Epoch 1/25...progress:8.99%....step:340/3780... Discriminator Loss: 2.9366... Generator Loss: 0.0905
Epoch 1/25...progress:9.26%....step:350/3780... Discriminator Loss: 2.6793... Generator Loss: 0.1169
Epoch 1/25...progress:9.52%....step:360/3780... Discriminator Loss: 1.4350... Generator Loss: 1.2931
Epoch 1/25...progress:9.79%....step:370/3780... Discriminator Loss: 2.2492... Generator Loss: 0.1752
Epoch 1/25...progress:10.05%....step:380/3780... Discriminator Loss: 0.8406... Generator Loss: 1.8311
Epoch 1/25...progress:10.32%....step:390/3780... Discriminator Loss: 1.1476... Generator Loss: 1.6161
Epoch 1/25...progress:10.58%....step:400/3780... Discriminator Loss: 0.8782... Generator Loss: 1.3749

Epoch 1/25...progress:10.85%....step:410/3780... Discriminator Loss: 0.9440... Generator Loss: 2.3991
Epoch 1/25...progress:11.11%....step:420/3780... Discriminator Loss: 0.8168... Generator Loss: 2.0876
Epoch 1/25...progress:11.38%....step:430/3780... Discriminator Loss: 0.8455... Generator Loss: 1.0280
Epoch 1/25...progress:11.64%....step:440/3780... Discriminator Loss: 2.2358... Generator Loss: 0.1792
Epoch 1/25...progress:11.90%....step:450/3780... Discriminator Loss: 2.1000... Generator Loss: 0.2152
Epoch 1/25...progress:12.17%....step:460/3780... Discriminator Loss: 0.7925... Generator Loss: 1.4258
Epoch 1/25...progress:12.43%....step:470/3780... Discriminator Loss: 2.1785... Generator Loss: 1.2142
Epoch 1/25...progress:12.70%....step:480/3780... Discriminator Loss: 0.8376... Generator Loss: 1.5873
Epoch 1/25...progress:12.96%....step:490/3780... Discriminator Loss: 0.8985... Generator Loss: 2.5251
Epoch 1/25...progress:13.23%....step:500/3780... Discriminator Loss: 0.8211... Generator Loss: 1.6519
Epoch 1/25...progress:13.49%....step:510/3780... Discriminator Loss: 0.8854... Generator Loss: 2.8512
Epoch 1/25...progress:13.76%....step:520/3780... Discriminator Loss: 0.8860... Generator Loss: 1.3317
Epoch 1/25...progress:14.02%....step:530/3780... Discriminator Loss: 1.4568... Generator Loss: 2.9661
Epoch 1/25...progress:14.29%....step:540/3780... Discriminator Loss: 1.1235... Generator Loss: 0.8529
Epoch 1/25...progress:14.55%....step:550/3780... Discriminator Loss: 1.9129... Generator Loss: 0.2500
Epoch 1/25...progress:14.81%....step:560/3780... Discriminator Loss: 0.7975... Generator Loss: 1.4410
Epoch 1/25...progress:15.08%....step:570/3780... Discriminator Loss: 0.8428... Generator Loss: 1.2874
Epoch 1/25...progress:15.34%....step:580/3780... Discriminator Loss: 1.7809... Generator Loss: 1.1281
Epoch 1/25...progress:15.61%....step:590/3780... Discriminator Loss: 1.4543... Generator Loss: 0.4929
Epoch 1/25...progress:15.87%....step:600/3780... Discriminator Loss: 1.2894... Generator Loss: 0.5546

Epoch 1/25...progress:16.14%....step:610/3780... Discriminator Loss: 0.9234... Generator Loss: 2.4267
Epoch 1/25...progress:16.40%....step:620/3780... Discriminator Loss: 2.8460... Generator Loss: 0.0922
Epoch 1/25...progress:16.67%....step:630/3780... Discriminator Loss: 1.1839... Generator Loss: 0.6369
Epoch 1/25...progress:16.93%....step:640/3780... Discriminator Loss: 1.6901... Generator Loss: 0.3321
Epoch 1/25...progress:17.20%....step:650/3780... Discriminator Loss: 1.4853... Generator Loss: 0.4731
Epoch 1/25...progress:17.46%....step:660/3780... Discriminator Loss: 1.7443... Generator Loss: 0.3586
Epoch 1/25...progress:17.72%....step:670/3780... Discriminator Loss: 0.9020... Generator Loss: 1.6566
Epoch 1/25...progress:17.99%....step:680/3780... Discriminator Loss: 1.6857... Generator Loss: 1.4632
Epoch 1/25...progress:18.25%....step:690/3780... Discriminator Loss: 0.8633... Generator Loss: 1.7218
Epoch 1/25...progress:18.52%....step:700/3780... Discriminator Loss: 1.8965... Generator Loss: 1.1640
Epoch 1/25...progress:18.78%....step:710/3780... Discriminator Loss: 0.8089... Generator Loss: 1.2937
Epoch 1/25...progress:19.05%....step:720/3780... Discriminator Loss: 2.0558... Generator Loss: 1.3412
Epoch 1/25...progress:19.31%....step:730/3780... Discriminator Loss: 1.9959... Generator Loss: 0.2520
Epoch 1/25...progress:19.58%....step:740/3780... Discriminator Loss: 1.0697... Generator Loss: 1.1000
Epoch 1/25...progress:19.84%....step:750/3780... Discriminator Loss: 1.3090... Generator Loss: 0.9893
Epoch 1/25...progress:20.11%....step:760/3780... Discriminator Loss: 1.7281... Generator Loss: 1.0718
Epoch 1/25...progress:20.37%....step:770/3780... Discriminator Loss: 0.9096... Generator Loss: 1.0252
Epoch 1/25...progress:20.63%....step:780/3780... Discriminator Loss: 1.1363... Generator Loss: 1.3904
Epoch 1/25...progress:20.90%....step:790/3780... Discriminator Loss: 0.8160... Generator Loss: 1.0757
Epoch 1/25...progress:21.16%....step:800/3780... Discriminator Loss: 0.9474... Generator Loss: 1.0572

Epoch 1/25...progress:21.43%....step:810/3780... Discriminator Loss: 1.7853... Generator Loss: 0.3171
Epoch 1/25...progress:21.69%....step:820/3780... Discriminator Loss: 0.8858... Generator Loss: 1.0936
Epoch 1/25...progress:21.96%....step:830/3780... Discriminator Loss: 0.6532... Generator Loss: 2.6104
Epoch 1/25...progress:22.22%....step:840/3780... Discriminator Loss: 1.5463... Generator Loss: 1.3285
Epoch 1/25...progress:22.49%....step:850/3780... Discriminator Loss: 0.7612... Generator Loss: 1.4031
Epoch 1/25...progress:22.75%....step:860/3780... Discriminator Loss: 1.6828... Generator Loss: 0.3768
Epoch 1/25...progress:23.02%....step:870/3780... Discriminator Loss: 1.1383... Generator Loss: 2.6189
Epoch 1/25...progress:23.28%....step:880/3780... Discriminator Loss: 0.9578... Generator Loss: 1.2157
Epoch 1/25...progress:23.54%....step:890/3780... Discriminator Loss: 1.1439... Generator Loss: 0.6777
Epoch 1/25...progress:23.81%....step:900/3780... Discriminator Loss: 1.7220... Generator Loss: 0.7275
Epoch 1/25...progress:24.07%....step:910/3780... Discriminator Loss: 0.8163... Generator Loss: 1.0294
Epoch 1/25...progress:24.34%....step:920/3780... Discriminator Loss: 1.4606... Generator Loss: 0.4576
Epoch 1/25...progress:24.60%....step:930/3780... Discriminator Loss: 1.8830... Generator Loss: 0.2629
Epoch 1/25...progress:24.87%....step:940/3780... Discriminator Loss: 1.3100... Generator Loss: 0.5509
Epoch 1/25...progress:25.13%....step:950/3780... Discriminator Loss: 1.0080... Generator Loss: 0.8271
Epoch 1/25...progress:25.40%....step:960/3780... Discriminator Loss: 1.4915... Generator Loss: 0.6623
Epoch 1/25...progress:25.66%....step:970/3780... Discriminator Loss: 1.2976... Generator Loss: 1.3785
Epoch 1/25...progress:25.93%....step:980/3780... Discriminator Loss: 1.9297... Generator Loss: 0.2493
Epoch 1/25...progress:26.19%....step:990/3780... Discriminator Loss: 1.2237... Generator Loss: 0.6136
Epoch 1/25...progress:26.46%....step:1000/3780... Discriminator Loss: 1.6167... Generator Loss: 0.8827

Epoch 1/25...progress:26.72%....step:1010/3780... Discriminator Loss: 0.8298... Generator Loss: 1.2450
Epoch 1/25...progress:26.98%....step:1020/3780... Discriminator Loss: 1.8139... Generator Loss: 0.2783
Epoch 1/25...progress:27.25%....step:1030/3780... Discriminator Loss: 1.8288... Generator Loss: 0.2788
Epoch 1/25...progress:27.51%....step:1040/3780... Discriminator Loss: 0.7185... Generator Loss: 1.2765
Epoch 1/25...progress:27.78%....step:1050/3780... Discriminator Loss: 0.7038... Generator Loss: 1.9092
Epoch 1/25...progress:28.04%....step:1060/3780... Discriminator Loss: 1.6458... Generator Loss: 0.3796
Epoch 1/25...progress:28.31%....step:1070/3780... Discriminator Loss: 1.6736... Generator Loss: 0.3247
Epoch 1/25...progress:28.57%....step:1080/3780... Discriminator Loss: 0.9654... Generator Loss: 0.8919
Epoch 1/25...progress:28.84%....step:1090/3780... Discriminator Loss: 1.9084... Generator Loss: 0.2589
Epoch 1/25...progress:29.10%....step:1100/3780... Discriminator Loss: 1.0863... Generator Loss: 1.5886
Epoch 1/25...progress:29.37%....step:1110/3780... Discriminator Loss: 1.0400... Generator Loss: 1.2753
Epoch 1/25...progress:29.63%....step:1120/3780... Discriminator Loss: 1.5819... Generator Loss: 1.3215
Epoch 1/25...progress:29.89%....step:1130/3780... Discriminator Loss: 1.0943... Generator Loss: 0.8743
Epoch 1/25...progress:30.16%....step:1140/3780... Discriminator Loss: 0.9881... Generator Loss: 0.8101
Epoch 1/25...progress:30.42%....step:1150/3780... Discriminator Loss: 1.8561... Generator Loss: 1.7873
Epoch 1/25...progress:30.69%....step:1160/3780... Discriminator Loss: 1.0489... Generator Loss: 0.7961
Epoch 1/25...progress:30.95%....step:1170/3780... Discriminator Loss: 1.2886... Generator Loss: 1.6620
Epoch 1/25...progress:31.22%....step:1180/3780... Discriminator Loss: 1.2187... Generator Loss: 0.5895
Epoch 1/25...progress:31.48%....step:1190/3780... Discriminator Loss: 1.3218... Generator Loss: 0.5196
Epoch 1/25...progress:31.75%....step:1200/3780... Discriminator Loss: 1.5088... Generator Loss: 0.4184

Epoch 1/25...progress:32.01%....step:1210/3780... Discriminator Loss: 1.1240... Generator Loss: 0.6717
Epoch 1/25...progress:32.28%....step:1220/3780... Discriminator Loss: 1.0239... Generator Loss: 1.1555
Epoch 1/25...progress:32.54%....step:1230/3780... Discriminator Loss: 1.6955... Generator Loss: 0.3233
Epoch 1/25...progress:32.80%....step:1240/3780... Discriminator Loss: 0.9675... Generator Loss: 1.3141
Epoch 1/25...progress:33.07%....step:1250/3780... Discriminator Loss: 1.6199... Generator Loss: 0.3793
Epoch 1/25...progress:33.33%....step:1260/3780... Discriminator Loss: 0.7565... Generator Loss: 1.3829
Epoch 1/25...progress:33.60%....step:1270/3780... Discriminator Loss: 1.6385... Generator Loss: 0.3699
Epoch 1/25...progress:33.86%....step:1280/3780... Discriminator Loss: 1.6718... Generator Loss: 0.3544
Epoch 1/25...progress:34.13%....step:1290/3780... Discriminator Loss: 1.4263... Generator Loss: 0.4690
Epoch 1/25...progress:34.39%....step:1300/3780... Discriminator Loss: 1.0944... Generator Loss: 1.0372
Epoch 1/25...progress:34.66%....step:1310/3780... Discriminator Loss: 1.7420... Generator Loss: 0.2995
Epoch 1/25...progress:34.92%....step:1320/3780... Discriminator Loss: 0.6552... Generator Loss: 1.8058
Epoch 1/25...progress:35.19%....step:1330/3780... Discriminator Loss: 0.8485... Generator Loss: 2.8014
Epoch 1/25...progress:35.45%....step:1340/3780... Discriminator Loss: 1.3084... Generator Loss: 0.6164
Epoch 1/25...progress:35.71%....step:1350/3780... Discriminator Loss: 1.4216... Generator Loss: 1.0545
Epoch 1/25...progress:35.98%....step:1360/3780... Discriminator Loss: 1.0022... Generator Loss: 0.9091
Epoch 1/25...progress:36.24%....step:1370/3780... Discriminator Loss: 1.5746... Generator Loss: 0.4262
Epoch 1/25...progress:36.51%....step:1380/3780... Discriminator Loss: 0.9101... Generator Loss: 0.9384
Epoch 1/25...progress:36.77%....step:1390/3780... Discriminator Loss: 0.6990... Generator Loss: 1.6440
Epoch 1/25...progress:37.04%....step:1400/3780... Discriminator Loss: 1.7239... Generator Loss: 1.1849

Epoch 1/25...progress:37.30%....step:1410/3780... Discriminator Loss: 1.1416... Generator Loss: 0.7930
Epoch 1/25...progress:37.57%....step:1420/3780... Discriminator Loss: 1.2138... Generator Loss: 2.2894
Epoch 1/25...progress:37.83%....step:1430/3780... Discriminator Loss: 1.1276... Generator Loss: 1.3225
Epoch 1/25...progress:38.10%....step:1440/3780... Discriminator Loss: 1.7343... Generator Loss: 0.3159
Epoch 1/25...progress:38.36%....step:1450/3780... Discriminator Loss: 1.2907... Generator Loss: 0.7114
Epoch 1/25...progress:38.62%....step:1460/3780... Discriminator Loss: 1.5751... Generator Loss: 0.4393
Epoch 1/25...progress:38.89%....step:1470/3780... Discriminator Loss: 0.9115... Generator Loss: 1.4774
Epoch 1/25...progress:39.15%....step:1480/3780... Discriminator Loss: 1.2858... Generator Loss: 0.5638
Epoch 1/25...progress:39.42%....step:1490/3780... Discriminator Loss: 1.2207... Generator Loss: 0.8748
Epoch 1/25...progress:39.68%....step:1500/3780... Discriminator Loss: 1.6012... Generator Loss: 0.3907
Epoch 1/25...progress:39.95%....step:1510/3780... Discriminator Loss: 1.5707... Generator Loss: 2.1960
Epoch 1/25...progress:40.21%....step:1520/3780... Discriminator Loss: 1.2181... Generator Loss: 0.7975
Epoch 1/25...progress:40.48%....step:1530/3780... Discriminator Loss: 0.9906... Generator Loss: 0.8674
Epoch 1/25...progress:40.74%....step:1540/3780... Discriminator Loss: 1.3559... Generator Loss: 0.6329
Epoch 1/25...progress:41.01%....step:1550/3780... Discriminator Loss: 1.8038... Generator Loss: 0.2924
Epoch 1/25...progress:41.27%....step:1560/3780... Discriminator Loss: 1.1420... Generator Loss: 0.8023
Epoch 1/25...progress:41.53%....step:1570/3780... Discriminator Loss: 1.2450... Generator Loss: 0.5492
Epoch 1/25...progress:41.80%....step:1580/3780... Discriminator Loss: 1.4304... Generator Loss: 0.8767
Epoch 1/25...progress:42.06%....step:1590/3780... Discriminator Loss: 1.2620... Generator Loss: 1.0086
Epoch 1/25...progress:42.33%....step:1600/3780... Discriminator Loss: 1.5485... Generator Loss: 0.4498

Epoch 1/25...progress:42.59%....step:1610/3780... Discriminator Loss: 0.9065... Generator Loss: 1.0810
Epoch 1/25...progress:42.86%....step:1620/3780... Discriminator Loss: 1.6616... Generator Loss: 0.3443
Epoch 1/25...progress:43.12%....step:1630/3780... Discriminator Loss: 1.6801... Generator Loss: 0.3365
Epoch 1/25...progress:43.39%....step:1640/3780... Discriminator Loss: 1.4040... Generator Loss: 0.4979
Epoch 1/25...progress:43.65%....step:1650/3780... Discriminator Loss: 0.9746... Generator Loss: 0.9584
Epoch 1/25...progress:43.92%....step:1660/3780... Discriminator Loss: 1.8837... Generator Loss: 0.2569
Epoch 1/25...progress:44.18%....step:1670/3780... Discriminator Loss: 0.9480... Generator Loss: 0.8874
Epoch 1/25...progress:44.44%....step:1680/3780... Discriminator Loss: 1.2285... Generator Loss: 1.1653
Epoch 1/25...progress:44.71%....step:1690/3780... Discriminator Loss: 1.5255... Generator Loss: 0.4187
Epoch 1/25...progress:44.97%....step:1700/3780... Discriminator Loss: 2.1071... Generator Loss: 0.2027
Epoch 1/25...progress:45.24%....step:1710/3780... Discriminator Loss: 1.8211... Generator Loss: 0.3210
Epoch 1/25...progress:45.50%....step:1720/3780... Discriminator Loss: 1.5588... Generator Loss: 0.3843
Epoch 1/25...progress:45.77%....step:1730/3780... Discriminator Loss: 1.7141... Generator Loss: 0.3155
Epoch 1/25...progress:46.03%....step:1740/3780... Discriminator Loss: 0.8290... Generator Loss: 1.4615
Epoch 1/25...progress:46.30%....step:1750/3780... Discriminator Loss: 1.4105... Generator Loss: 0.5058
Epoch 1/25...progress:46.56%....step:1760/3780... Discriminator Loss: 1.6617... Generator Loss: 2.3673
Epoch 1/25...progress:46.83%....step:1770/3780... Discriminator Loss: 1.1455... Generator Loss: 0.7535
Epoch 1/25...progress:47.09%....step:1780/3780... Discriminator Loss: 1.3628... Generator Loss: 0.5126
Epoch 1/25...progress:47.35%....step:1790/3780... Discriminator Loss: 0.7332... Generator Loss: 2.1539
Epoch 1/25...progress:47.62%....step:1800/3780... Discriminator Loss: 0.9984... Generator Loss: 0.9496

Submitting This Project¶

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.

print("Done!")

Done!